Nanofiltration Reduces Salt in Distributed Water
The NF plant started to operate in September 1997, and was in full operation in January 1998. During the winter of 1997-98, the Clifton Water District did not need to buy any water from the city of Grand Junction, but instead sold water to it.
Pretreatment
The river water is treated with alum, 1 ppm Catfloc LS, before flowing into one of several retention ponds. During summer, powder activated carbon (PAC) is added to remove taste from the water when biological activity is high. Periodically one of the retention ponds is emptied to remove the settled solids.
After settling, the river water goes to a rapid 35-45 mesh sand filter. The alum dosage before the settling ponds is controlled from streaming current measurement of the water after the sand filter.
Chlorine is added in gas form after the sand filters, and the chlorinated filtrate is stored in a clearwell. Fluoride is added in the form of H2F6Si in the clearwell at a location which is close to the pumps for water distribution, but away from the pumps in the same clearwell which pump the filtrate to the NF plant.
The NF plant consists of 4 skids in parallel. Each skid has its own two multimedia filters in parallel followed by cartridge filters, with sodium bisulfite, sulfuric acid and antiscalant addition between. The multimedia filters have six layers of four different materials: anthracite, manganese greensand, gravel and garnet.
The antiscalant Flocon 260 from FMC controls sulfate scaling. In the pilot plant testing performed in fall 1996, there was a severe membrane flux loss, which was determined to be from barium sulfate scaling.
Addition of 93 percent sulfuric acid solution to a pH of 6.9-7.0 prevents calcium carbonate precipitation.
NF UNITS
There are four practically identical NF units, named skids 1-4. They operate at nominal 85 percent permeate recovery.
It is a three stage system arranged in a 9-5-3 array, with six 8-inch diameter by 40-inch long elements with polyamide nanofiltration membranes in each housing, totaling 102 elements in each of the four skids. The stainless steel housings of the four port design minimize the required plumbing. The permeate is taken out from both ends of each housing.
Most of the operational data are transferred electronically to a computer, which among other things normalizes the permeate flow rate. The computer monitor will on demand show many different parameters, including the actual operating conditions and the values of different operational parameters between specified dates.
The permeate goes directly to a blending tank, without the help of a transfer pump. This contributes to 17 psig permeate pressure in the permeate manifold during operation of the NF skids.
Mode of Operation
The permeates from the four NF skids flow in a single pipe to a 1 million gallon tank, where it is blended into filtered river water for distribution.
The water level in the permeate tank determines how many NF skids are in operation, according to this schedule:
When an NF unit is not operating, it is flushed with NF permeate for 4 minutes daily. If it will not be in operation for a week, it is preserved with a solution of 1 percent sodium bisulfite. The preservative solution is changed every two weeks.
The installation of the NF units started in September 1997, and ended in January 1998.
Plant Performance
Skid 1 was in full operation in late October 1997 and so far is the only skid with performance data from the same full month in two different years.
The permeate flow rate for Skid 1 was 400-460 gallons per minute (gpm) during the first winter, and first week of operation upon restart in August. Then it gradually decreased to stay at about 320 gpm. In January 1998 at 2°C, feed pressure of 184 psig was required to produce 410 gpm permeate. At the same low temperature in December 1998, Skid 1 produced 314 gpm permeate at 165 psig feed pressure. During the first week after restart in August, the permeate flow rate was 460 gpm at 121 psig and 22°C.
The permeate flow rate from a membrane element strongly depends on pressure and temperature, but is slightly dependent on the solutes concentrations. In order to evaluate the performance of the membrane elements, the data must be normalized by calculation of the fouling factor (FF). The fouling factor is multiplied with the water permeability of the nominal element, which together with measured pressures, solute concentrations, feed flow rate, and temperature, give the measured permeate flow rate of the skid. Thus, a fouling factor of 1.0 means that the elements operate with normal flux performance, and a fouling factor of 0.80 means that the elements on average have lost 20 percent of their nominal flux performance.
With Skid 1, the fouling factor decreased to 0.67 in August, and improved somewhat to 0.80 after cleaning. A cleaning in November brought the fouling factor up to 1, which is the nominal flux performance. The fouling factor decreased to 0.87 during another 300 hours of operation, after which it again was restored to 1 by cleaning.
If the pressure in December '98 had been increased from 165 to 184 psig, as during the low temperatures the previous year, the permeate flow rate would have exceeded 400 gpm after cleaning. However, there was no need for a higher permeate flow rate, and the fouling rate would most likely have increased with increasing permeate flow rate.
Summary
The nanofiltration plant at the city of Clifton met the requirements during its first year in operation. Membrane fouling results in a cleaning frequency, which most of the time exceeds once monthly. The cause of membrane fouling still is being investigated.